The Challenge
New private equity ownership brought a fresh board and heightened expectations for a specialty health and wellness retailer sitting on two decades of customer data. Leadership suspected the customer base had fundamentally shifted but lacked the infrastructure to prove it or act on it. Marketing teams could not build effective lookalike audiences or deliver personalized experiences for digital acquisition campaigns. January represents a make-or-break month for annual performance, and the company needed actionable customer segments activated before Q1. The incoming board expected meaningful, novel insights beyond what internal teams had previously surfaced, creating a clear deadline with executive visibility.
The Constraint
Customer data lived across 40+ source tables spanning CRM, transactions, loyalty programs, email (Responsys), and SMS (Attentive) with no unified data model. Analysts spent days manually wrangling data for each analysis request. The existing analytics infrastructure relied on one-off Excel analyses and PowerPoint decks accumulated over two decades and could not support the segmentation sophistication required for modern digital marketing. The analytics team lacked self-serve capabilities, creating bottlenecks that blocked rapid experimentation and delayed critical marketing decisions.
The Approach
We embedded alongside the analytics team and built a bronze/silver/gold medallion architecture in Snowflake using dbt, consolidating 40+ source tables into a unified customer dimension with 43 attributes. We applied behavioral clustering to the high-value base and pivoted the primary analysis metric from revenue to gross margin dollars based on early client feedback, ensuring targeting of truly profitable customers.
The Outcome
7% of customers drive 35% of total revenue. Segmentation revealed five distinct male-dominated high-value segments, confirming a fundamental shift away from the traditional female customer profile and validating leadership's hypothesis with data. Time-to-insight dropped from days to hours by eliminating manual data wrangling. A small but high-value 'Web-Only Female Shopper' segment (2% of the high-value base) was identified as lost to competitors, establishing a clear win-back priority with defined targeting parameters.
What This Unlocked
Ready-to-activate audiences for Google, Meta, and TikTok, built directly from validated behavioral segments.
85/15 budget split strategy allocating spend between known high-value targeting and prospecting, now informing every marketing investment decision.
Executive validation secured. Insights are actively shaping the company's growth and rebranding strategy.
Modular star schema architecture functions as reusable building blocks. The team iterates on new analyses without rebuilding from scratch.
Future integrations already in the pipeline: Insider.com for web and app engagement, Power BI for automated reporting.
Services
Tech Stack
Results
Customer records unified
5.6 million from 40+ sources
High-value segments identified
5 distinct behavioral clusters
Time to executive presentation
90 days from kickoff
Self-serve analytics
Enabled for analyst team